[Bug target/124133] [14/15/16 Regression] ICE in decompose, at wide-int.h:1049 on powerpc64le-linux-gnu since r14-4793

[jakub at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124133

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |meissner at gcc dot gnu.org,
                   |                            |ppalka at gcc dot gnu.org,
                   |                            |segher at gcc dot gnu.org
     Ever confirmed|0                           |1
            Summary|[16 Regression] ICE in      |[14/15/16 Regression] ICE
                   |decompose, at               |in decompose, at
                   |wide-int.h:1049 on          |wide-int.h:1049 on
                   |powerpc64le-linux-gnu       |powerpc64le-linux-gnu since
                   |                            |r14-4793
   Last reconfirmed|2026-02-16 00:00:00         |2026-02-17
             Status|UNCONFIRMED                 |NEW
           Priority|P3                          |P2

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Started with r14-4793-gdad311874ac3b3cf4eca1c04f67cae80c953f7b8
I'd say with the departure of NON_DEPENDENT_EXPR it is highly undesirable at
least for C++ to call fold_convert and fold_build2 in the
targetm.resolve_overloaded_builtin hook, I'd think it should convert or build1
NOP_EXPR or something similar instead.